Bölümler

  • AI Return on Optionality - with Dhiraj Rajaram, Founder, Mu Sigma
    19 Nov 2026

    Many large enterprises are spending heavily on AI and still not getting the returns they expected. Dhiraj Rajaram’s view is that the problem is less about the technology and more about how organizations solve problems, remember context, and allocate capital.

    In this episode, Ray Rike talks with Dhiraj Rajaram, Founder, Chairman, and CEO of Mu Sigma, one of India’s first unicorns and today one of its most profitable AI companies, with more than 140 Fortune 500 clients. Mu Sigma raised about $20 million of primary capital and returned about $900 million to investors. Dhiraj shares his framework for why enterprise AI is underdelivering, why context and memory matter more than more data, and why CFOs should look beyond ROI to what he calls return on options.

    In this episode:

    • Learning matters more than knowing: As the cost of knowledge and intelligence moves toward zero, human judgment and intuition become more valuable. Dhiraj explains why companies that use AI to replace people will underperform the ones that invest in how people and AI work together.


    • Why enterprise AI is not delivering: Databases capture the action trace, meaning what happened. They miss the perception and decision traces that explain why and how it happened. As headcount falls and institutional knowledge disappears, AI programs lose the context they need to work.


    • Building a field of context: Mu Sigma uses program and project ontologies to capture an organization’s memory of purpose and memory of process. These ontologies turn business problems into factors, metrics and a hypothesis matrix that sits alongside the existing data.


    • Stop overspending on clean data: Dhiraj argues that clean, connected data is an unattainable goal. He suggests putting roughly half of the data budget into data quality and a meaningful share into building context. That requires the CIO to work much more closely with the CFO, COO and functional leaders.


    • From ROI to return on options: In a volatile, uncertain world, optionality is the answer to uncertainty. Dhiraj makes the case that CFOs should allocate AI resources based on a mix of ROI and return on options, and explains why that requires CFOs to lead the dialogue with their boards.


    • Build a better kitchen: Solving specific problems with a new technology is not enough. Organizations that also rethink how they solve problems will capture the most value from AI.


    Follow the AI to ROI podcast on your favorite podcast app, leave a five-star rating, and subscribe to the AI to ROI newsletter at ai2roi.substack.com. To suggest a future guest, reach out to Ray Rike on LinkedIn.

    See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

    Daha fazlasını göster Daha Az Göster
    29 dk.
  • The Great AI Trade + Blue Wave Mash-Up
    19 Nov 2026

    Last Tuesday’s election sent a clear message to the AI trade. Data centers are now a mainstream political issue, and voters from both parties want local control over how and where AI infrastructure gets built.

    In this Big Story edition of AI to ROI, Ray Rike and Peter Buchanan look at the election results, how the markets reacted, and what changes for the 2028 capacity outlook. They also cover what enterprise buyers and boards should do about vendor risk, pricing and infrastructure planning over the next 6 to 18 months.

    In this episode:

    • The voter mandate: 56% of voters in an NBC exit poll called data centers a major or crucial factor in their vote. 46 of 47 local ballot measures passed or are on track to pass, and at least 17 are outright bans on large data centers. Governors who pledged to stop or pause construction now oversee states holding roughly 40% of US data center capacity.


    • Why local government holds the levers: Federal permitting reform may pass in the lame duck session, but cities and townships still control zoning and property taxes. Ray and Peter score the odds of federal and state policy outcomes and explain why enterprises should plan for a patchwork of state rules.


    • The market repricing: The Nasdaq 100 fell 6.4%, CoreWeave bond yields rose above 16%, and Microsoft tightened oversight of its neo cloud partners. The market is repricing counterparty risk, and AI demand is holding up.


    • 2027 is locked and 2028 is at risk: UBS estimates one third of the computing power due online in 2028 is at risk. Texas alone could account for up to 30% of US gigawatt capacity in 2028, and Ray and Peter walk through base, upside and downside cases for that state.


    • Winners and losers: Hyperscalers with energized, permitted sites and more than $2 trillion in contracted backlog are well positioned. Anthropic (about $900 billion in commitments against $146 billion in cash) and OpenAI (which raised $11 billion against a $30 billion target) face a harder road. So do neo clouds that still need funding.


    • The board checklist: Treat local permitting politics as a supply risk, plan for state-by-state rules, ask your AI vendors who funds them, and budget for higher prices.

    Subscribe to the AI to ROI newsletter at ai2roi.substack.com for the full analysis, including the probability scorecard and the Texas scenario analysis. New Big Story editions publish every Tuesday, Metric of the Week every Wednesday, and News and Analysis every Friday.

    See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

    Daha fazlasını göster Daha Az Göster
    38 dk.
  • The Perilous State of the AI Trade
    12 Nov 2026

    In this Big Story episode of the AI to ROI podcast, Ray Rike, Founder and CEO of Benchmarkit, and Peter Buchanan, Managing Partner at NewPlan, look at the side of the AI trade that most executives don’t see until something breaks: the debt. Hyperscalers, Neoclouds, and third-party funders are paying for the AI buildout with more borrowed money and more off-balance-sheet structures than ever before.

    At the same time, Bain, Gartner, and Theory Ventures each identify a large gap between projected AI revenue and what it will take to pay that debt back.

    Ray and Peter explain what the numbers mean for C-suite executives, board members, and enterprise buyers who rely on AI vendors.

    Key topics discussed:

    • Same chips, different cost of money: Two operators can sell access to the same NVIDIA chips and borrow at 7% and 15%. That spread is about $80 million a year on $1 billion in annual loan repayments. CoreWeave’s debt insurance costs implied default odds above 60% within five years, and tenant quality is now setting the price of capital.


    • Who pays for the buildout: The top five hyperscalers will spend $780 billion on data center capex in 2026, almost five times the 2023 level. Free cash flow no longer covers the spend, so new AI-related bonds and loans reached $489 billion through September 2026, against $322 billion in all of 2025. Special purpose vehicles like Meta’s Hyperion campus with Blue Owl, along with construction delays, add risk that doesn’t show up on the balance sheet.


    • The payback gap: Bain identifies a $4.2 trillion revenue gap. Gartner estimates AI providers need $8.5 trillion in cumulative revenue through 2030 to earn a 12% return, against a forecast of no more than $3.6 trillion. Tomasz Tunguz estimates AI revenue has to grow from roughly $150 billion today to $1.2 to $1.5 trillion a year by 2030 to service the debt.


    • What Ramp’s spending data shows: Real corporate card and expense data points to three cracks in the bull case: paid adoption is stagnating, blended token prices are down 41% from their peak, and the top 1% of AI-using firms drive roughly 80% of OpenAI and Anthropic B2B revenue. Enterprise buyers are acting like cost-conscious software buyers.


    • Neoclouds, Oracle, and three scenarios: CoreWeave’s top three customers produced 72% of second-quarter revenue. Nscale has only about 2.5% of its $103.4 billion in contracted value live. Oracle spent $1.23 for every dollar of operating cash last quarter. Ray and Peter walk through three ways the market could play out, with a flight to quality and consolidation among second-tier providers as the most likely.


    • What enterprise buyers should do now: Underwrite AI vendors the way a lender would. Ask about anchor customers, how capacity is financed, and what sits off the balance sheet. Match model tiers to tasks, set budget caps, negotiate pricing that tracks falling token costs, and build exit options into contracts.


    This episode is analysis, not investment advice. For the full article with every figure and source, subscribe to the AI to ROI newsletter on Substack.

    If you are finding value in the AI to ROI podcast, please subscribe on your favorite podcast app, leave a five-star rating, and share it with a colleague.

    See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

    Daha fazlasını göster Daha Az Göster
    40 dk.
  • The Business-Ready AI Gap - with Christian Lund, Co-Founder, Templafy
    11 Nov 2026

    In this episode of the AI to ROI podcast, Ray Rike talks with Christian Lund, Co-founder of Templafy, about why widespread AI adoption is not yet turning into measurable business value.

    Christian walks through findings from Templafy’s Business-Ready AI Gap 2026 research, a survey of more than 2,000 knowledge workers in the US and UK combined with Templafy platform data. The research shows that 90% of workers use AI to create business documents at least weekly, yet 95% still edit the output before it is ready to use, and they spend almost four hours a week reviewing and correcting it.

    The conversation covers how CFOs and operating leaders should measure AI productivity, why company context has to be built into AI workflows, and what Christian learned rebuilding an established SaaS company for the AI era.

    Key topics discussed:

    • The business-ready gap: Why fast AI generation moves the work into editing and review, and why measuring productivity at the point of first draft overstates the gains. As Christian puts it, “you’re done when you’re done.”


    • From generation to orchestration: Only 30% of workers say approved templates and content are applied automatically by their AI tools, and 42% say their tools can’t get to approved company content. Christian explains why the business has to be the “third player” between the user and the model.


    • Agent-led document workflows: AI agent adoption among enabled Templafy users rose from 31% to 66% in nine months. The discussion covers where humans need to stay in the loop (setting direction and validating output) and how embedding rules into agents supports review and approval workflows in regulated industries.


    • Generic AI output as a business risk: 59% of respondents say AI-generated documents make companies sound alike, and 42% name loss of trust as a top risk. Christian argues that differentiation comes from putting brand, messaging, and positioning into the AI layer.


    • ROI guidance for CFOs: Start with specific use cases, compare against how the work was done before AI, and use the prompt data from your own users to direct AI toward business-critical outputs.


    • The founder’s view: How Templafy moved a cloud-era SaaS platform to AI-first, balanced risk-averse enterprise customers with early adopters, and used in-app, product-led distribution plus governance to drive adoption without adding risk.


    Christian closes with his career advice for recent graduates: be the person who sets direction for AI, or the person who validates its output, and ideally both.

    If you are finding value in the AI to ROI podcast, please subscribe on your favorite podcast app, leave a five-star rating, and connect with Ray Rike on LinkedIn to suggest future guests. To read the full Business-Ready AI Gap 2026 report, visit templafy.com.

    See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

    Daha fazlasını göster Daha Az Göster
    28 dk.
  • AI Governance as an Accelerator - with David Trier, CEO at ModelOp
    4 Nov 2026

    Enterprise AI spending keeps rising, yet most pilots never reach production and ROI stays hard to prove.

    Our host, Ray Rike talks with David Trier, CEO of ModelOp, about why AI governance speeds up production deployment when it is built as a repeatable, automated process. Dave has spent the last decade helping large enterprises put AI into production.

    Dave explains how leading companies define business impact before deployment, govern multi-agent systems at several levels, and manage AI spend as a portfolio of investments a CFO can track.

    Key Highlights

    • Why pilots stall before production: Each team builds its own path to production, so projects run into “enterprise friction” from legal, risk, compliance, security, data, and IT reviews. Dave’s fix is a set of enterprise-wide standards that every business unit follows, which still leaves departments free to choose their own tools and use cases. One customer used this approach with automated workflows and cut time from idea to production from 9 to 12 months to less than half that.


    • Governance runs for the life of the system: AI is probabilistic and depends on its data, so governance continues after launch. That means continuous risk monitoring, ongoing testing against verified data sets, traceability into execution logs, and a clear record of every model, tool, skill, and orchestrator version running.


    • Measure value before you deploy: Companies that get the most from AI answer three questions up front. What business impact are we targeting? Which metrics will show it? Do we have the instrumentation to track those metrics in production?


    • An AI system of record with one accountable owner: A typical AI solution involves 8 to 10 stakeholders. Dave argues for a single system of record that sits above every vendor’s agent platform (ERP, CRM, supply chain) and shows usage, risk, cost, and value in one place. Each business use case has one named AI owner who is accountable for it, even when 50 agents run underneath.


    • Manage AI as a portfolio, and drop token maxing: Dave’s advice to CFOs is to manage AI investments like a stock portfolio, using consistent impact metrics and a central view of cost and benefit. ROI ownership sits with the head of AI and the CIO, working with the CFO, who decides whether to continue or kill each initiative. He calls tracking each employee’s token consumption a vanity metric that boards should stop accepting.


    Subscribe to the AI to ROI podcast on your favorite podcast app, leave a five-star rating, and connect with Ray Rike on LinkedIn to suggest future guests.

    See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

    Daha fazlasını göster Daha Az Göster
    30 dk.
  • The Keys to AI-Native Software Company Hypergrowth
    3 Nov 2026

    AI-native software companies now treat $100M in ARR within 24 months as the hypergrowth benchmark.

    Ray Rike, Founder and CEO of Benchmarkit, and Peter Buchanan, Managing Partner of New Plan, discuss the October 20 AI to ROI newsletter edition.

    They look at how Lovable, Gamma, Cursor, ElevenLabs, and Legora reached $100M faster than any earlier generation of B2B software companies, and what that growth costs. They cover revenue-per-employee and gross-margin benchmarks, the role of forward-deployed engineers in enterprise land-and-expand, and what boards should track behind the growth headline.

    Key Highlights

    • The new hypergrowth threshold: Lovable reached $100M in ARR in about nine months with fewer than a dozen people and no full-time CMO, CRO, or CFO. Sierra, Cursor, ElevenLabs, and Legora all passed $100M in under 24 months, with a median of about 18 to 22 months. Ray and Peter discuss where to start the clock: at commercial launch or at the first $1M in revenue. They also note that 24 to 36 months of enterprise experimentation budgets drove high churn, followed by a growing “boomerang” of returning customers.


    • Revenue per employee is the efficiency signal: The median for private SaaS companies is $175K to $200K, and for public B2B SaaS it is about $400K, up from $300K two years ago. ICONIQ’s AI-native Pacesetters average $655K, and Bessemer’s Supernovas average $1.13M. Agentic coding, small rapid-iteration teams, and lean go-to-market functions drive the gap. Legora’s founder recalibrates every core function each quarter, not every year.


    • The gross margin trade-off: Building on frontier LLMs and hyperscaler infrastructure comes at the cost of gross margin. Bessemer’s Supernovas average about 25% gross margin, and the AI-native companies Ray works with average 40% to 50%, well below the traditional SaaS range of 70% to 80%. Cursor reported 23% gross margin at the time of its acquisition by SpaceX. Leaders are improving unit economics by embedding open-weight models, matching each task to the right price-performance model, and building proprietary models, as Harvey did.


    • Forward-deployed engineers drive enterprise expansion: As AI-native companies move up market, forward-deployed engineers (FDEs) close product gaps, feed requirements back to product teams, and find expansion use cases. ElevenLabs grew its FDE team from 12 to more than 60 in about a year. The more complex the workflow and integrations, the more FDEs matter to time to value.


    • What boards should track: Look past the growth headline to the operating model: gross margin, which models sit under the product, how much value comes from FDEs instead of software, and funding runway. The four lessons for operators are to obsess over an easy-to-adopt product focused on a tight ideal customer profile, to build sales capacity before product-led growth stalls, to use FDEs to shorten time to value and expand accounts, and to be ready to scale headcount quickly once enterprise demand arrives.


    Subscribe to the AI to ROI podcast on your favorite podcast app, read the full Big Story edition at ai2roi.substack.com, and connect with Ray Rike and Peter Buchanan on LinkedIn.

    See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

    Daha fazlasını göster Daha Az Göster
    35 dk.
  • AI & Organization Design - a Return on Intelligence Conversation with Brian Solis, Head of Innovation ServiceNow
    27 Oct 2026

    Brian Solis, Head of Global Innovation at ServiceNow, joins Ray Rike to explain why AI strategy is an organizational design challenge, not simply a technology deployment.

    The conversation moves beyond making today's work faster to ask what companies could do differently with AI. Brian and Ray discuss the gap between AI ambition and execution, why end-to-end workflows need redesign, and how leaders can turn individual productivity gains into measurable business value.

    What Ray and Brian Covered

    • Redesign the work before adding agents. Customer journeys cross functions, but systems, data, budgets, and metrics often do not. Brian explains why connecting people and processes is part of the work, alongside governance, accountability, security, and observability.
    • AI maturity is more than adoption. Drawing on ServiceNow's Enterprise AI Maturity Index, Brian discusses the distance between experimenting with AI and changing how work flows across the enterprise.
    • Iteration and innovation need different goals. Improving today's work can free capacity, but the return depends on what happens next. Brian's two-mode approach pairs efficiency gains with deliberate investment in new work, customer relationships, and growth.
    • Measure outcomes, not activity. Ray and Brian connect return on intelligence to revenue per employee, output relative to human effort, and the economic value created when people and agents work together.
    • Leadership starts with better questions. Brian shares lessons from his work with executives and his book, Infinite Company, on learning, unlearning, and designing organizations around outcomes rather than the traditional org chart.

    If you are finding value in these conversations, please subscribe on your favorite podcast app, leave a five-star rating, and connect with Ray Rike on LinkedIn to suggest future guests.



    See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

    Daha fazlasını göster Daha Az Göster
    32 dk.
  • The Quest to Create an Auditable, Trackable, and Accurate AI Budget
    24 Oct 2026

    Companies have AI spend tracking in place, but very few can forecast it. In this Big Story episode of AI to ROI, Ray Rike and Peter Buchanan walk through the October 13 newsletter, “The Quest to Create an Auditable, Trackable, and Accurate AI Budget.” The discussion draws on Benchmarkit’s research with Mavvrik covering 396 organizations, which was recently cited by the Wall Street Journal and Apollo.

    The headline finding is that tracking is not the same as forecasting. Nearly every company tracks AI infrastructure costs, yet only 11% forecast AI spend within 10% accuracy, and more than half miss by over 26%. The miss affects the business directly. 62% of respondents say unexpected AI costs materially changed a business decision in the past year, about one in four stopped an initiative, and 48% of companies that embed LLMs in their products had to raise prices.

    Ray and Peter cover the budget overruns at Uber, Rippling, and Amazon, along with the shared pathologies behind them: misaligned incentives like token maxing, usage that went unforecasted and untracked, and no measurable ROI. They also share results from their own model test.

    A model that cost $0.15 per attempt ended up costing almost $57 per completed task once failures, retries, and human rework were included.

    Key Takeaways

    • Tracking is not visibility: Most companies see spend at a summary level, but only about a third track the cost of the agents they have in production, and 46% have no real-time alerting for overruns.
    • Six cost layers: A full AI budget covers tokens and inference, AI infrastructure, agents, AI management applications, labor, and the growing verification tax.
    • Price the finished task, not the attempt: Completion rates, retry multipliers, and human rework drive total cost far more than list token prices do.
    • A six-step CFO framework: Calculate the full cost of a finished task, trace every dollar to an owner through chargebacks, set guardrails and automatic stops, govern workflows rather than people, and audit invoices every quarter.
    • The verification tax is real: In a survey of 6,000 senior executives, 90% saw no measurable AI productivity impact, in large part because of the labor needed to review and correct AI output.

    The bottom line: measure AI cost at the most granular level possible, and always tie that cost to the economic value of the outcome.

    See Privacy Policy at https://art19.com/privacy and California Privacy Notice at https://art19.com/privacy#do-not-sell-my-info.

    Daha fazlasını göster Daha Az Göster
    36 dk.